Recent clinical results are supportive that induction of differentiation is an additional knowledge of the process of terminal differentiation and the mechanism of action of inducers. Studies were conducted on the metabolism of retinoic acid (RA), a potent inducer of differentiation. The human myeloid leukemia cell line, HL60, has been a useful model system for studying terminal differentiation. Although many biological effects of RA have been described, the mechanism for these actions is unknown. We have extended our discovery that a covalent bond is formed between RA and protein in many cell lines. In HL60 cells there are at least 20 retinoylated proteins which are primarily localized in the nucleus. Based on chemical property, the RA moiety is probably linked to protein via a thioester bond. We found that proteins in undifferentiated and squamous-differentiated normal human epidermal keratinocytes were retinoylated after treatment with [3H] RA. The major retinoylated proteins were identified as cytokeratins based on their profile in two-dimensional gel electrophoresis and their immunoreactivity with anti-keratin monoclonal antibodies. In contrast to HL60 proteins, the covalently bound [3H]RA in keratinocytes was not removed by mild hydrolysis with methanolic-KOH indicating that RA is not linked to the cytokeratins by a thioester bond. Retinoylated proteins also were found in MCF-7 breast tumor cell lines. The apparent widespread modification of proteins by RA prompted us to examine if other hormones of the steroid/thyroid superfamily also covalently bind to protein. In either HL60 cells and MCF-7 cell lines or both we found that 17beta-estradiol, progesterone, 1,25(OH)2 vitamin D3, triiodothyronine, thyroxine, and prostaglandin E2 covalently bind to proteins. These results raise the possibility that some activities of ligands of the steroid hormone receptor family and of prostaglandin. E2 may be mediated by their covalent modification of proteins.